Utilization of "Intersection" Methodology for Optimization with Many Objectives in Designed Experiments of Material Processing

Material processing involves many factors, and evaluation of final quality of a product also relates to many indexes from different respects. Thus the optimization of qualities and processing of a product is an optimization problem with many objectives (MOO) inevitably. Although there are some approaches proposed to deal with optimization problem with many objectives in nowadays, the inherent shortcomings in these approaches make them puzzled, which include their missing standpoint and use of additive algorithm containing subjective factors. Currently, an "intersection" methodology for optimization with many objectives is proposed by initiating a novel idea of favorable probability to depict the favorite degree of an alternative in optimal option impersonally, which aims to characterize the concurrent optimization of many objectives in a system in spirits of probability theory and set theory. In this article, some regulations are put forward for performing optimal option of material processing parameters in designed experiments of response surface methodology, orthogonal experiment design and uniform experiment design by means of the total / global favorable probability. In the treatment, the total / global favorable probability of an alternative is the decisive indicator in the optimal option uniquely, which transfers the optimization problem with many objectives into a mono-objective one. The result indicates that the novel approach can be employed to deal with the optimal problem of designed experiments for material processing rationally.